Noncompliance in hemodialysis: predictors and survival analysis

Abstract

Noncompliance with hemodialysis (HD), depending on the definition, occurs in 2% to more than 50% of patients. To better understand predictors and outcomes of noncompliance, we evaluated patient characteristics associated with noncompliance and the impact of noncompliance on survival. Using data from two USRDS special studies, we identified 6,251 patients who were on dialysis for more than 1 year for inclusion in this study. Noncompliance was defined in four ways: skipping one or more HD sessions in a month, shortening by 10 or more minutes one or more HD sessions in a month, an interdialytic weight gain (IWG) of more than 5.7% of dry weight, or a serum phosphate (PO4) of greater than 7.5 mg/dL. Sociodemographic predictors of noncompliance were identified using logistic regression. Survival analysis was done using Cox proportional hazards models with adjustments for sociodemographics, comorbid conditions, and dose of HD. Overall, 8.5% of patients skipped HD, 20% shortened HD (7% three or more times), 10% had more than a 5.7% IWG, and 22% had a PO4 greater than 7.5. There was a significant correlation among the measures of noncompliance. Blacks (adjusted odds ratio [AOR] = 2.10), patients aged 20 to 39 years (AOR = 1.62), and smokers (AOR = 1.34) were significantly more likely to skip HD than whites, patients aged 40 to 59 years, and nonsmokers, respectively (P < 0.01 for each). Similar results were seen for the other measures of noncompliance, except for PO4, in which blacks were significantly less likely to be noncompliant (NC) (AOR = 0.85, P < 0.05). Compared with compliant patients, those who skipped one or more HD sessions in a month had a 25% higher risk of death (P < 0.01). Those who had greater than a 5.7% IWG had a 35% higher risk of death (P < 0.001), whereas those with a PO4 > 7.5 had a 13% higher risk of death (P < 0.05). Overall, patients who shortened HD sessions did not have a higher risk of death, but those who shortened three or more in 1 month had a 20% higher risk of death (P < 0.05). Compliance with a medical regimen is a complex issue. Noncompliance in HD often, but not always, is associated with a higher risk of an adverse outcome. This is a US government work. There are no restrictions on its use.